Dept title


Current state of knowledge on Highly Pathogenic Avian Influenza

This chapter provides background information on avian influenza viruses and the factors that have led to the emergence, spread and persistence of HPAI viruses. It uses experiences before and after the emergence of Asian-lineage H5N1 HPAI viruses as a guide.

2.1 Avian influenza viruses

Aquatic birds are the natural hosts of type A influenza viruses. Influenza viruses are categorized by the glycoprotein spikes on their surface – the ‘HA’ or haemagglutinin protein (16 different subtypes identified so far) and the ‘NA’ or neuraminidase protein (nine subtypes). All 16 HA and nine NA subtypes (see box on Terminology) have been detected in various combinations in aquatic wild birds, and for the most part these viruses live harmoniously with their natural hosts, establishing a short-lived subclinical enteric infection (Webster and Hulse, 2004). Occasionally, viruses cross from aquatic wild birds to poultry or mammals, and new genotypes of virus may become established in these non-natural or spillover hosts. Once adapted to non-avian hosts, these viruses may lose the capacity to multiply in the gut of aquatic birds (Kobasa et al, 2001)

Of the influenza viruses that cross to terrestrial poultry, the most important are those of the H5 and H7 subtypes. These have the capacity to mutate into HPAI viruses that multiply systemically in chickens, often causing very high mortality in infected flocks (Alexander, 2000).

Other influenza virus subtypes have also crossed into terrestrial poultry and have established in these birds. Among these, the H9N2 subtype is the most frequently detected, and has spread across Asia and into the Middle East over the past 15 years (Alexander, 2000). This subtype is a low pathogenicity virus but, in combination with other pathogens, can cause severe respiratory disease in poultry (Kishida et al, 2004). H9N2 viruses are now circulating in pig populations in China (Chen et al, 2006b) and have been associated with cases of clinical disease in this species (Peiris et al, 2001). They have also caused non-fatal infections in humans (Butt et al,2005).


Throughout this document, the term ‘Asian-lineage H5N1 HPAI’ is used to describe highly pathogenic avian influenza viruses of the H5N1 subtype linked to the line of viruses first detected in geese in Guangdong province in China in 1996. ‘H5N1 HPAI’ is used to describe the disease caused by these viruses. Although multiple genetic variants of these viruses have emerged, they still form a lineage distinct from other H5N1 viruses. All HPAI viruses of the H5N1 subtype detected in the past 11 years belong to this lineage.

he term HPAI has a very specific meaning, and relates to the ability of the virus to cause disease in experimentally inoculated chickens (i.e. its virulence). It does not reflect the capacity of these viruses to produce disease in other species.

The term ‘H5N1 HPAI’ is used rather than just ‘H5N1’ to avoid confusion with unrelated low pathogenicity avian influenza (LPAI) viruses of the H5N1 subtype that have been detected in poultry and wild birds.

Influenza viruses have a segmented genome and the capacity to undergo gene reassortment. Theoretically, this process could occur whenever two different influenza viruses co-infect the same cell. In addition, minor changes to individual genes occur relatively frequently, leading to changes in these genes over time. Some cases of gene recombination have also occurred, most notably in outbreaks of HPAI in Chile (2002) and Canada (2004).

Avian influenza viruses are named according to their type (all avian influenza viruses are Type A), subtype (i.e. by the HA and NA glycoproteins they possess, which is written as H5N1, H7N7, etc.), by the species of animal from which they are isolated, by the geographic location from which they were isolated (often country, province or state level), and by laboratory reference number and year of isolation. (e.g. A/Goose/Guangdong/1/96 [H5N1]).

Nomenclature beyond this is not fully standardized; no internationally-agreed system has been used for naming different genetic variants within subtypes. For example, numerous new genotypes of Asian-lineage H5NI HPAI viruses have emerged through reassortment. Although these viruses all retain a version of the “parent” H5 gene, they were initially subdivided on the basis of the makeup of other genes coding for internal proteins.

he first reported Asian-lineage H5N1 HPAI virus ­­– A/Goose/Guangdong//1/96 – is used as a starting point for comparisons of genotypes (Chen et al, 2004; Guan et al, 2004). As these (or similar) viruses evolved, they acquired genes from other influenza viruses and the genotypes formed were named by using letters. Researchers at the University of Hong Kong were very active in this area and the letters they used (e.g. the ‘Z’ genotype) have been widely adopted. Unfortunately this system has not been applied consistently by all researchers.

This variation in genotype does not necessarily capture information on variations in the HA genes. This is covered by a different system of nomenclature in which multiple clades (lineages) and subclades (sublineages) of H5N1 virus haemagglutinin genes have been identified. The World Health Organization (WHO) initially suggested standardizing the nomenclature (using clades and subclades to group these viruses) in 2005 and 2006 (WHO, 2005; WHO, 2006). This has since been further revised (WHO, 2007) with at least ten clades and many subclades among the Asian-lineage H5N1 HPAI viruses currently recognized.

Because the clades and subclades are based on the genetic relationships between HA genes, individual clades may contain different genotypes (the latter is determined by the constellation of gene coding for internal proteins, not the HA gene). Therefore, use of any given system of nomenclature does not capture all of the information on the changes that are occurring. This can make comparison of molecular data difficult, especially if different labels are used by unrelated research groups to describe similar viruses or even the same strains of virus.

2.1.1 Background information on HPAI Early cases of HPAI

HPAI is not a new disease (Alexander, 1987). Its occurrence predates the industrialization of the poultry subsector. 

Although interpretation of historical data on HPAI (before the so-called ‘virological era’) is complicated by the potential for confusion of this disease with other diseases such as Newcastle disease, HPAI was first identified in the late 19th century and probably occurred earlier (Alexander, 1987). At least one early outbreak in 1901 spread across international borders through movement of poultry (Wilkinson and Waterson, 1975), providing forewarning of the challenges to be faced when trade in poultry and poultry products would become globalized later in the 20th century. In the United States of America in the 1920s, HPAI viruses were apparently spread via movement of poultry by rail and through live-bird markets (Alexander, 1987).

The precise origins of these early viruses and the events that facilitated the mutation of the causative viruses from low to high pathogenicity are not known (see section 2.1.2). It has been suggested that infection with HPAI was endemic in Germany, Italy and Egypt for a number of years in the late 19th and/or the first half of the 20th century (Alexander, 1987). This indicates that, in the past, not all outbreaks of the disease were rapidly contained, in contrast to those from the 1950s to the 1970s, which were generally isolated cases that were controlled rapidly. For example, the 1959 outbreak in the United Kingdom occurred on an isolated farm and resulted in the death of virtually all poultry, conditions which reduced the likelihood of spread (Alexander, 1987). Recent modelling studies support the proposition that low farm density can result in limited or even no onward transmission of infection (Truscott et al, 2007).

The rise in the number of cases of HPAI

There has been an increase in the number of reported outbreaks of HPAI in the past eight years (Capua and Alexander, 2004), even excluding recent outbreaks of H5N1 HPAI in Asia and elsewhere. Those associated with HPAI viruses (other than the H5N1 subtype) have occurred in the Americas (United States of America, Chile and Canada), Europe (Italy and the Netherlands) and Pakistan.

It is still not entirely clear why this increase occurred. A number of factors have been proposed, including the increase in the global poultry population, the increase in intensification of the poultry industry with a concurrent increase in free-range production, and even climate change leading to alterations in wild bird migratory paths (Capua and Alexander, 2004). None of these has been proved to be the cause of this increase, which is likely to be multifactorial.

Improved diagnostic capability and surveillance may have contributed to some of this increase. Enhanced surveillance has allowed detection of cases that might otherwise have gone undiagnosed, such as the clinically mild 'outbreak' in Texas in 2004 (Lee et al, 2005).

Much of the increase in importance of HPAI stems from the severe effects of the cases that did occur. Three large outbreaks in Italy (1999), the Netherlands (2003) and Canada (2004) involving areas of high poultry density led to the destruction of tens of millions of poultry, including many healthy birds (Capua et al, 2003; Power, 2005; Stegemen et al, 2004).

Source of viruses for outbreaks of HPAI

Given the level of attention paid to HPAI recently, it is sobering to reflect on how little has been demonstrated about the mode of introduction and spread of the disease. Wild aquatic birds have been proposed as the most likely source of LPAI viruses that converted to HPAI viruses in many HPAI outbreaks (excluding, for the moment, those caused by Asian-lineage H5N1 HPAI viruses) (see for example Alexander, 2007) but the evidence, in most cases, is circumstantial. Table 1 provides a summary of some of these outbreaks of HPAI, and includes information on both the possible source of the virus and factors contributing to the emergence of the highly pathogenic strain. These cases demonstrate that the precise source of infection in most outbreaks was not identified.

The lack of information on source(s) also applies to outbreaks involving Asian-lineage H5N1 HPAI viruses. Few in-depth investigations have been conducted on disease outbreaks caused by these viruses, especially in developing countries, and even when these have been performed, the route of entry or source of virus has not been proved. For example, tracing of contacts and movements was performed in Japan (Nishiguchi et al, 2005) and the Republic of Korea (Wee et al, 2006) following outbreaks in 2003-04, but even after these detailed studies, the authors could only speculate on how the virus entered these countries. Similarly, recent epidemiological studies on outbreaks in Israel (2006) and the United Kingdom (2007) were not able to prove how the virus was introduced (Balicer et al, 2007; Defra, 2007).

Table 1. Selected HPAI outbreaks – possible sources and environmental factors

Virus subtype
Possible contributing factors
Australia 1975-76
Dam on farm attracted wild birds
Untreated drinking water used for poultry
Substandard biosecurity (Turner, 1981)
Australia 1985
Dam on farm attracted wild birds
Untreated drinking water used for poultry
Substandard biosecurity (Westbury, 1997)
Australia 1992
Adjacent to free-range duck farm
Substandard biosecurity (Westbury, 1997)
Australia 1994
Many wild birds on watercourse due to drought
Untreated drinking water used for poultry (Westbury, 1997)
Australia 1997
Wild birds in vicinity
Free-range farmed emu chicks infected subclinically on one of three infected farms
Dead bird pick-up vehicle travelled to multiple farms including the infected farms (Selleck et al, 2003)
Canada 2004
Migratory birds in area prior to the outbreak (Power, 2005)
Italy 1997-98
Marketing of infected birds
Rearing of birds in the open
Presence of mixed species (Capua et al, 2003)
Italy 1999-2000
Original source unknown – presumably wild bird
Widespread infection with low pathogenicity virus prior to emergence of highly pathogenic strain (Capua et al,2003)
Mexico 1994
Widespread circulation of low pathogenicity viruses prior to emergence of highly pathogenic strain (Villarreal, 2006)
Netherlands 2003
Unknown but likely to be wild bird introduction
LPAI virus similar to the HPAI virus detected in wild mallards (Munster et al, 2005)
USA 1983-84
Unknown but likely to be wild bird introduction
Infection of poultry with low pathogenicity viruses
Live–bird markets (Suarez and Senne, 2000)
USA 2004
Link to live-bird markets
Similar LPAI viruses circulating for some time prior to emergence of HPAI strain (Lee et al, 2005)

An association between wild birds and infection with LPAI viruses in farmed turkeys in Minnesota, United States, has long been recognized (Halvorson et al, 1985). It is also considered a probable risk factor for repeated outbreaks of both LPAI and HPAI in northern Italy (Capua et al, 2003), along with the high density of poultry and the large number of live birds imported to the area.

In Italy, H7N3 LPAI viruses virtually identical to those found in wild birds in 2001 were subsequently detected in farmed turkeys in 2002-03 (Campitelli et al, 2004). A close relationship has been demonstrated between H5 and H7 LPAI viruses isolated from wild mallards and those from outbreaks of HPAI in Europe, including the 2003 H7N7 outbreak in the Netherlands (Munster et al, 2005). The origin of the 2004 outbreak of HPAI in Canada is not known, but this and other cases developed in poultry farms located in areas with seasonally high populations of migratory wild birds.

Outbreaks of HPAI in Australia from the 1970s to the 1990s were on farms implementing inadequate biosecurity measures. This included several cases in which untreated drinking water from ponds or rivers frequented by wild waterbirds was supplied to poultry (Westbury, 1997). It is pertinent to note that no new cases of HPAI have been diagnosed in Australia in the past ten years, which is probably attributable (at least in part) to implementation of enhanced biosecurity measures on commercial farms. These were introduced in response to earlier cases of HPAI, and to a major outbreak of Newcastle disease on large intensive farms. It may also reflect a reduction in the number of wild birds due to prolonged drought (Turner, 2004).

2.1. Mutation from LPAI virus to HPAI virus

A key event in the genesis of all HPAI viruses is conversion (mutation) of an H5 or H7 LPAI virus to an HPAI virus. This has occurred in the past following multiplication of LPAI viruses of these subtypes in chickens but it is not known whether this is an essential prerequisite.

Virulence of avian influenza viruses is a polygenic trait (Suarez et al, 2004) that usually results from insertion or substitution of multiple basic amino acids at the cleavage site of the HA protein. These are not normally present in LPAI viruses. This mutation allows the HA protein to be cleaved by a broad range of proteases, allowing the viruses to multiply systemically (Alexander, 2000). Other novel mechanisms for conversion of LPAI viruses to HPAI viruses have been described in outbreaks of HPAI in Chile (2002) and Canada (2004). These arose through recombination between the HA gene and that of another gene coding for an internal protein, leading to insertion of additional amino acids at the HA cleavage site (Suarez et al, 2004; Pasick et al, 2005). Modification of the cleavage site appears to be an essential condition, but is not the only factor that determines virulence (Londt et al, 2007).

Even though the molecular events surrounding mutation from an LPAI virus to an HPAI virus are known, the factors that lead to this mutation are not clear for many outbreaks of H5 and H7 avian influenza viruses, including the first of the Asian-lineage H5N1 HPAI viruses.

An HPAI virus has been generated experimentally by repeat passage of a LPAI virus through chickens by air sac and intracerebral inoculation (Ito et al, 2001) but the exact triggers for this change under natural conditions are not known. In some earlier outbreaks of HPAI, it was evident that the change from a LPAI virus to an HPAI virus followed introduction of LPAI virus to large flocks of commercial poultry. This change apparently occurred within a matter of days in some outbreaks (as was the case of the 2004 Canadian outbreak [Bowes et al, 2004]). On the other hand, in some Central American countries, low pathogenicity H5N2 strains have circulated in poultry for a number of years without developing into highly pathogenic strains. Even in Mexico, where mutation of a LPAI H5N2 virus to an HPAI virus occurred in 1994 and this HPAI virus strain was subsequently eliminated, H5N2 LPAI viruses continue to circulate (Villarreal, 2006) but have not reverted to high pathogenicity.

The Asian-lineage H5N1 HPAI viruses differ somewhat from those in earlier HPAI outbreaks in that the HPAI viruses first detected in 1996 were not eliminated, and have circulated in highly pathogenic form for over 11 years (Sims et al, 2005). These viruses have evolved considerably over this time, but no closely related low pathogenic precursor strains of the H5 subtype have been isolated, (Duan et al, 2007; Mukhtar et al, 2007). Presumably, such a virus existed prior to 1996 but has never been detected or reported (Sims et al, 2005). The HA and NA genes in LPAI viruses most closely related to those in Goose/Guangdong/1/96 were isolated experimentally from aquatic birds in Japan (Mukhtar et al, 2007).

All subsequent Asian-lineage H5N1 avian influenza viruses have remained highly pathogenic and their origins can be traced back to viruses similar to those found in geese in Guangdong in 1996 (Sims et al, 2005). Even those viruses that have infected domestic ducks subclinically in many parts of Asia retain the multiple basic amino acids at the cleavage site and are highly pathogenic for terrestrial poultry (Chen et al, 2004).

No evidence has been provided to indicate that the first Asian-lineage H5N1 HPAI virus emerged in an intensive poultry farm. If this conversion from LPAI to HPAI occurred in geese (the type of poultry from which it was first isolated), this would not have involved industrialized production facilities because geese in southern China were then reared mainly in small flocks with only a few large semi-intensive production units.

The lack of markers of adaptation to chickens (e.g. a deletion in the stalk of the NA glycoprotein (Matrosovich et al, 1999) in the original 1996 strains of H5N1 HPAI virus also suggests, but does not prove, that limited circulation of this virus occurred in chickens before 1997. A stalk deletion in the NA was not detected in viruses of this lineage until March 1997, when it was detected in a virus isolated from a dead chicken on a farm in the Hong Kong Special Administration Region (SAR) (Bender et al, 1999).

These observations, coupled with evidence that other HPAI viruses emerged in the early 20th century prior to intensification of the poultry industry, indicate that the circulation of a LPAI virus in industrialized poultry rearing systems, although considered an important factor in the emergence of some HPAI strains, is not an essential prerequisite for the genesis of an HPAI virus.